Proc. Natl. Acad. Sci. USAVol. 80, pp. 5075-5079, August 1983Medical Sciences

Parasite sequestration in Plasmodium falciparum malaria: Spleenand antibody modulation of cytoadherence ofinfected erythrocytes

(monkey/endothelium/surface antigen/cell binding)

PETER H. DAVID*, MARCEL HOMMELt, Louis H. MILLER* IROKA J. UDEINYA* ANDLYNETTE D. OLIGINOt*Laboratory of Parasitic Diseases, National Institutes of Health, Bethesda, Maryland 20205; and tDivision of Tropical Medicine, Brigham and Women's Hospital andHarvard Medical School, Boston, Massachusetts 02115

Communicated by Richard M. Krause, April 11, 1983

ABSTRACT Sequestration, the adherence of infected eryth-rocytes containing late developmental stages of the parasite (tro-phozoites and schizonts) to the endothelium of capillaries and ven-ules, is characteristic of Plasmodium falciparum infections. Wehave studied two host factors, the spleen and antibody, that in-fluence sequestration of P. falciparum in the squirrel monkey. Se-questration of trophozoite/schizont-infected erythrocytes that oc-curs in intact animals is reduced in splenectomized animals; in vi-tro, when infected blood is incubated with monolayers of humanmelanoma cells, trophozoite/schizont-infected erythrocytes fromintact animals but not from splenectomized animals bind to themelanoma cells. The switch in cytoadherence characteristics ofthe infected erythrocytes from nonbinding to binding occurs witha cloned parasite. Immune serum can inhibit and reverse in vitrobinding to melanoma cells of infected erythrocytes from intact an-imals. Similarly, antibody can reverse in vivo sequestration as shownby the appearance of trophozoite/schizont-infected erythrocytesin the peripheral blood of an intact animal after inoculation withimmune serum. These results indicate that the spleen modulatesthe expression of parasite alterations of the infected erythrocytemembrane responsible for sequestration and suggest that the pre-vention and reversal of sequestration could be one of the effectormechanisms involved in antibody-mediated protection against P.faleiparum malaria.

Among the four species of malaria infecting man, one distin-guishing feature of Plasmodium falciparum is that only eryth-rocytes containing young forms of the parasite (rings) are foundin the peripheral blood (1); erythrocytes containing more ma-ture forms (trophozoites and schizonts) are sequestered in thepostcapillary venules of various organs where they adhere toendothelial cells (2). The mechanisms of parasite sequestrationremain unclear. Ultrastructural studies have suggested that ad-herence of parasitized erythrocytes to the vascular endothe-lium occurs by means of excrescences on the infected eryth-rocyte membrane that have been called knobs (3, 4); antigenspresent on these knobs are recognized by immune serum of thehost (5, 6). Cytoadherence of infected erythrocytes can be stud-ied in vitro through the binding of infected erythrocytes tomonolayers of human endothelial cells (7); infected erythro-cytes also bind to monolayers of human melanoma cells, whichcan replace endothelial cells for in vitro studies of parasite se-questration (8).

Sequestration favors the development of the parasite by pro-tecting it from the filtering action of the spleen (9); it is alsoresponsible for the severe clinical forms of cerebraL malaria in

which brain capillaries are obstructed by sequestered parasites.Means of inhibiting sequestration may therefore hinder para-site development and alleviate the clinical severity of the dis-ease.

In the P. falciparum infection of the squirrel monkey, wehave studied two host factors that influence both sequestrationin vivo and adherence to melanoma cells in vitro-namely, thespleen and antibody. First, by comparing the cytoadherenceproperties of infected erythrocytes in nonsplenectomized andin splenectomized monkeys infected with the same strain of theparasite, we have shown that, as a result of splenectomy, se-questration in vivo is reduced and that infected erythrocytesfrom these splenectomized animals no longer bind to mela-noma cells in vitro. Second, the passive transfer of immune seruminto an infected intact animal can reverse sequestration; thiscorrelates with the fact that antibody can inhibit and reversebinding of infected erythrocytes to melanoma cells in vitro.

MATERIALS AND METHODSAnimals. Female Bolivian squirrel monkeys (Saimiri sci-

ureus) (South American Primates, Miami, FL) were usedthroughout this study. Nonsplenectomized and splenectomizedanimals will be referred to as S' and S animals, respectively.

Parasites. The Ugandan Palo Alto/PLF-3 strain of P. falci-parum adapted to the squirrel monkey (10) and the derived clone,PLF-3/BLL (11), were used. In view of the differences ob-served in ref. 11 between parasites maintained in S+ and S-animals, we maintained two different lines by serial transfer ineither S+ or S animals.

Animals were infected by intravenous injection of infectedblood. Parasites were obtained by venipuncture of anesthe-tized infected animals. Infected blood was cryopreserved andthawed according to the method described of Diggs et al. (12).Short-term culture of fresh or thawed infected blood was car-ried out as described by Trager and Jensen (13) except that fetalbovine serum was- substituted for human serum.Immune Sera. Immune serum 31/S+ was obtained in the

following manner:. A S+ animal was infected with parasites froma S+ animal; after self-cure of this infection, the animal waschallenged three times by intravenous inoculation of infectedblood from a S+ animal; none of these challenges were followedby detectable-parasitemia. Serum was collected 7 days after thelast challenge. Immune serum 30/S- was obtained in a similar

Abbreviations: S+, nonsplenectomized; S-, splenectomized; MCBA,melanoma cell-binding assay; TC medium, tissue culture medium.t Present address: Dept. of Parasitology, Liverpool School of TropicalMedicine, Pembroke P1., Liverpool, L3 5QA, United Kingdom.

5075

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertise-nent" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 3,

202

0

5076 Medical Sciences: David etalP

fashion from a S- animal infected and challenged three timeswith parasites from a S animal.

Tissue Culture Medium. The tissue culture (TC) mediumused for both the culture of melanoma cells and malaria par-asites was RPMI 1640 medium containing glutamine and so-dium bicarbonate (M A Bioproducts, Walkersville, MD), 25 mMHepes, gentamicin at 0.1 mg/ml, and 10% heat-inactivated fe-tal bovine serum (Flow Laboratories).Melanoma Cells. Amelanotic human melanoma cells (Amer-

ican Type Culture Collection no. CRL 1585, designation C32r)were put into culture as in ref. 8, with the following modifi-cations. Cells were suspended in TC medium at a concentrationof 2 x 105/ml; 0.5 ml of this cell suspension was layered on a2.2-mm glass coverslip in a 35-mm plastic Petri dish. Cells wereallowed to settle and stick to the coverslip overnight at 370C inhumidified 5% C02/95% air; 1.5 ml of TC medium was thenadded. Binding assays were carried out 24-48 hr after platingof the melanoma cells.Melanoma Cell-Binding Assay (MCBA). When P. falcipa-

rum-infected blood is incubated with monolayers of humanmelanoma cells, trophozoite/ and schizont-infected erythro-cytes specifically bind to these cells; there is no binding of un-infected erythrocytes or of erythrocytes infected with rings (8).To more readily quantify the results of this type of MCBA, welabeled the erythrocytes with 51Cr and modified the assay de-scribed in ref. 8 in the following manner: Infected blood wascultured for 24 hr to allow partial maturation of the ring stages.Parasitemia was adjusted to 2-5% by addition of normal Saimirierythrocytes; cultures were then centrifuged and the pellet wassuspended to a 5% cell suspension in TC medium. One mil-licurie of sodium chromate (New England Nuclear; 1 mCi/mlin sterile saline; 1 Ci = 37 GBq) was added to 4 ml of cell sus-pension. After 1 hr at 370C in humidified 5% C02/95% air, thelabeled erythrocytes were washed three times with TC mediumand the pellet was suspended to a 20% suspension in TC me-dium. Medium was aspirated from the melanoma cell-contain-ing Petri dishes and replaced with 1 ml of fresh TC medium towhich 100 ,ul of the labeled erythrocyte suspension was added.Dishes were incubated at 37°C in 5% C02/95% air for 1 hr ona rocking platform (25 cycles per min); coverslips were then re-moved, washed by gentle dipping in five successive baths of TC

15

0

x

E

10 1

5

01 2 3 4

medium, and put into fresh Petri dishes containing 1 ml of TCmedium supplemented with 100 ,ul of 5% Nonidet P-40; themedium containing the 51Cr freed from the lysed erythrocytesbound to the melanoma cells was transferred into counting vialsand radioactivity was measured with a Packard 5230 gammacounter.

As shown in Fig. 1, binding measured in the MCBA cor-related with parasite maturation; ring-infected erythrocytes didnot bind to the melanoma cells, binding appeared at the tro-phozoite stage, and increased with parasite maturation from thetrophozoite stage (6,000 cpm) to the schizont stage (12,000 cpm).

Treatment of Erythrocytes with Enzymes. Erythrocytes(5 X 108; 4% parasitemia) were incubated in 1 ml of TC medi-um without serum containing different concentrations of tryp-sin (Sigma), chymotrypsin (Worthington), or neuraminidase(GIBCO) for 30 min at 37°C. Cells were washed three timeswith 10 ml of TC medium. Trypsin-treated cells were then in-cubated for 15 min at 37°C with soybean trypsin inhibitor (Mil-lipore; 1 mg/ml), and chymotrypsin-treated cells were incu-bated with phenylmethylsulfonyl fluoride (Sigma; 2 mM); cellswere then washed three times with TC medium. After enzymetreatment, erythrocytes were either labeled and used in theMCBA or cultured for 24 hr before labeling and MCBA.

Electron Microscopy. Erythrocytes infected with clone B-11taken from a S - or a S+ squirrel monkeywere cultivated in vitrofor 28 hr to obtain an optimal level of schizont-infected cells.The cell suspension was fixed for 15 min on ice in an equal vol-ume of Karnovsky's aldehyde fixative and further processed asdescribed (14). Stained thin sections were examined in a JEOL100C electron microscope.

RESULTSDifferences in the Degree of in Vivo Parasite Sequestration

in S+ and S Animals as Reflected by Peripheral Blood Smears.The differences between the evolution of parasitemia in a S+and a S animal are shown in Fig. 2. Daily blood smears fromthe S+ animal showed that mainly ring forms were present inthe peripheral blood, reflecting a high level of parasite se-questration; the parasitemia evolved in a sawtooth manner, de-

10

8

6

4

2

FIG. 1. Effect of parasite development on binding of infected eryth-rocytes to melanoma cells in vitro. Bars: 1, normal (uninfected) eryth-rocytes; 2-4, parasitized blood from a S+ animal (6% parasitemia, syn-chronous infection) cultured for 6, 24, and 36 hr, respectively. The 6-hrsample was 100% rings, the 24-hr sample was 20% rings/77% tropho-zoites/3% schizonts, and the 36-hr sample was 4% rings/10% tropho-zoites/86% schizonts. Results are expressed as mean ± 2 SD release of51Cr from detergent-lysed erythrocytes bound to melanoma cells in du-plicate cultures.

0 4 6 8 10 12 14 16 0 13 15 17 19 21Days

FIG. 2. Evolution of parasitemia in S+ (A) and S- (B) monkeys. S+and S- monkeys were inoculated intravenously with a stabilized formof infected blood (S+ and S- line, respectively). Thin blood films weremade daily and the percentage of erythrocytes containing rings (o) ortrophozoite/schizonts () was determined.

Proc. Natl. Acad. Sci. USA 80 (1983)

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 3,

202

0

Proc. Nati Acad. Sci. USA 80 (1983) 5077

creasing every other day when the majority of parasites wereat the trophozoite/schizont stage and were thus sequestered inthe tissues. This contrasted with what could be seen in a S-animal in which, at any one time, a mixture of different de-velopmental stages could be found in the peripheral blood, in-dicating that parasite sequestration was absent or dramaticallyreduced; parasitemia increased in a continuous fashion and con-sisted of a mixture of both ring and trophozoite/schizont stages.Such patterns were consistently observed in other animals, al-though the length of the prepatent period varied greatly fromone animal to another.

Comparison of Infected Erythrocytes from S' and S- An-imals in the MCBA. Was the decreased sequestration observedin S- animals merely a consequence of the removal of a po-tential splenic reservoir for trophozoite/schizont-infectederythrocytes or was it related to an alteration of the cytoad-herence properties of the infected erythrocyte? We comparedthe in vitro binding to melanoma cells of infected erythrocytesfrom S' and S monkeys by using infected blood samples witha similar parasitemia and degree of maturation. As shown inFig. 3, only infected erythrocytes from S' hosts bound to mel-anoma cells. The absence of binding of infected erythrocytesfrom S animals was confirmed in 11 experiments involving fourdifferent S monkeys.

Evolution of in Vitro Binding Characteristics of the ClonedParasite PLF-3/B11 After Transfer from a S- into a S' Mon-key. Cloned PLF-3/B11 was used to explore if and when a clonedparasite could switch the in vitro cytoadherence properties ofinfected erythrocytes from nonbinding to binding. Blood froma S- monkey infected with clone PLF-3/B11 was inoculatedinto a S' monkey. Infected erythrocytes from the inoculum didnot bind. The MCBA was carried out 8 and 20 days after in-oculation. By day 8 after inoculation, the parasites had gonethrough four cycles of multiplication and the majority of par-asites had invaded the erytirocytes of the S+ host. At this time,there was no binding of infected erythrocytes in the MCBA. Incontrast, infected erythrocytes did bind in the MCBA on day20 and binding ability was maintained after inoculation into an-other S+ animal. Values obtained for binding of the cloned par-asite in the S+ and S- monkeys are shown in Fig. 4.

Ultrastructure of the Infected Erythrocyte Surface. Com-parison of clone B-11 in erythrocytes of S+ and S monkeysshowed that both populations had knobs on their surface, con-firming earlier studies carried out with the parent line (14).

15

0

x

0.

10 _

5

0 3 4S3 4

FIG. 4. Binding of cloned and uncloned parasites. Bars: 1, normalerythrocytes; 2, PLF-3/S+ (original uncloned parasite population in anintact animal; 6% parasitemia/>80% trophozoite/schizonts); 3, B-11/S- (cloned parasite population from a splenectomized animal; 4.1% par-asitemia/>80% trophozoite/schizonts); 3, B-11/S+ (cloned parasitepopulation in an intact animal after switch, 5.5% parasitemia/>80%trophozoite/schizonts).

Effect of Enzyme Treatment of Infected Blood on Resultsof the MCBA. After 24 hr in culture, when parasites from S+animals had developed from rings to mature trophozoites orschizonts, infected blood was treated with neuraminidase, chy-motrypsin, or trypsin. Treatment with trypsin at 1 ug/ml orchymotrypsin at 10 pkg/ml totally abolished binding to mela-noma cells. In contrast, treatment with neuraminidase at 25 units/ml increased binding of infected cells of the S+ line by 50%;it did not affect the lack of binding of normal erythrocytes, oferythrocytes infected with ring-stage parasites, or of erythro-cytes infected with schizonts from S- animals. To gain addi-tional information on the origin of the alterations of the infectederythrocyte membrane responsible for the binding, infectedblood was treated with trypsin, washed, and then cultured foran additional 24 hr to allow parasite maturation. When carriedout at this early stage, enzyme treatment did not inhibit bind-ing to melanoma cells of the treated cells 24 hr later.

Inhibition and Reversal of in Vitro Binding by ImmuneSerum. The binding assay of infected erythrocytes from S+ an-imals was carried out in the presence of various dilutions ofserum 31/S+ and serum 30/S-. As shown in Fig. 5, serum 31/S+ totally inhibited binding up to a dilution of 1:100; the in-

20 40 100 200 400 1,000 2,000 10,000Serum dilution

FIG. 3. Effect of splenectomy on binding of infected erythrocytes tomelanoma cells. Bars: 1, normal erythrocytes; 2, infected blood from aS- animal (4.5% parasitemia/>80% trophozoite/schizonts); 3, infectedblood from a S+ animal (5.2% parasitemia/>80% trophozoite/schiz-onts).

FIG. 5. Inhibition by immune serum of binding of infected eryth-rocytes to melanoma cells. Binding of infected erythrocytes from a S+animal (3% parasitemia/>80% trophozoite/schizonts) was carried outin the presence of increasing dilutions of serum 31/S+ (@) and serum30/S- (0). El, Binding ± 2 SD in the presence of normal serum.

TI-

5

o 4'-4

x3

2F

1

0

3

2

x

0.

I

0

1 2

Medical Sciences: David et aL

1 2

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 3,

202

0

5078 Medical Sciences: David et al.

hibitory effect was significant up to a dilution of 1:2,000. Withserum 30/S-, no more than 50% inhibition could be obtainedand binding was no longer inhibited at a dilution higher than1:100.We also studied the ability of serum 31/S' to reverse bind-

ing. Binding of infected erythrocytes to melanoma cells wasallowed to occur in the absence of immune serum; coverslipscoated with melanoma cells to which infected erythrocytes hadbound were then rinsed and incubated for 1 hr at 37TC with a1:20 dilution of serum 31/S' or normal serum. The results ofthree experiments are given in Table 1. Up to 99% of the bounderythrocytes were detached by immune serum while only 22%were detached in the controls. Inhibition and reversal of bind-ing were confirmed by using two other immune sera from dif-ferent S animals.

Reversal of in Vivo Parasite Sequestration by Inoculationof Immune Serum. Two S' animals were infected with para-sites from a S+ animal. When parasitemia reached 5 and 13%,respectively, each monkey received an intravenous injection of1.5 ml of serum 31/S+; parasitemia was monitored before andimmediately after serum transfer (times 0, 5, 15, 30, 60, and120 min). Before inoculation with immune serum, few tropho-zoite/ or schizont-infected erythrocytes were seen in the pe-ripheral blood, reflecting the parasite sequestration that occursin S+ animals. The number of trophozoite/schizonts in the pe-ripheral bloodstream increased sharply during the first half hourafter serum injection and then slowly decreased (Fig. 6). Twenty-four hours after serum transfer, parasitemia was drastically re-duced (from 5.0 to 0.3% and from 13.0 to 0.4%) and more than40% of the parasites present consisted of abnormal tropho-zoites and schizonts ("crisis forms"). This was followed by a de-crease of the parasitemia to less than the detectable level. Nei-ther the immediate effect on circulating trophozoites andschizonts nor the delayed effect on total parasitemia were ob-served in animals injected with normal serum.

DISCUSSIONWe have examined the effects of splenectomy and of antibodyon cytoadherence of infected erythrocytes and have begun todefine the nature of the alterations of the infected erythrocytesurface involved in sequestration.We have shown that P. falciparum-infected erythrocytes from

S+ and S squirrel monkeys differ in their ability to be se-questered in vivo and to bind to melanoma cells in vitro. In-fected erythrocytes from S+ animals bind massively to mela-noma cells, in contrast to infected erythrocytes from S- animals,which do not bind at all. This correlates with the decreased par-asite sequestration observed in S- animals. Such differences incytoadherence properties of infected erythrocytes from S+ andS- squirrel monkeys show that the spleen plays a major role inmodulating the erythrocyte surface alterations responsible forparasite sequestration.

Table 1. Reversal of in vitro binding of infected erythrocytes tomelanoma cells by immune serum 31/S+

Exp. 1 Exp. 2 Exp. 3 Mean ± 2 SD% detached withnormal serum 20 16 30 22 ± 7.0

% detached withserum 31/S+ 97 100 99 98 + 1.5

The MCBA was carried out in the absence of immune serum. After

80

40

4 2 20

O °80

4-4020

4 6 Lu

A

B

15 30 60 120Time, min

FIG. 6. Kinetics of release of trophozoite/schizont-infected eryth-rocytes in the peripheral bloodstream after passive transfer ofimmuneserum; 1.5 ml of serum 31/S+ was inoculated at time 0 into two S+ an-imals, A and B. The number of trophozoite/schizont-infected eryth-rocytes was counted in the peripheral blood.

The mechanisms by which the spleen could influence thecytoadherence properties of infected erythrocytes include (i)initial differences in erythrocytes between uninfected S+ andS animals, (ii) direct modification of the surface of infectederythrocytes by the spleen, (iii) selection of nonbinding mu-

tants, and (iv) induction of changes in expression in a clonedparasite population. Results obtained with clone PLF-3/B11show that, 8 days after the infection of a S+ animal with par-

asites from a S- animal, erythrocytes from the S+ host con-taining trophozoites or schizonts still do not bind in vitro. Thisexcludes the possibility that differences in cytoadherence in S+and in S animals are related to initial differences in the eryth-rocyte populations from the S+ and S monkeys; it also arguesagainst the hypothesis that the spleen secondarily modifies theparasite-induced alterations of the infected erythrocyte mem-brane responsible for sequestration. By day 20, infected eryth-rocytes bind in vitro; this switch in cytoadherence propertiessuggests that the spleen induces a change in parasite expressionon the surface of the infected erythrocyte. This is comparablewith what has been observed with surface antigens on Plas-modium knowlesi-infected erythrocytes; a parasite clone can losethe ability to express variant-specific erythrocyte surface an-tigens after passage into a splenectomized rhesus monkey (15).Few observations have been reported concerning the evo-

lution of P. falciparum in splenectomized human hosts. Sple-nectomy performed in a patient infected with P.falciparum hasbeen reported to have modified the pattern of parasitemia dur-ing a subsequent attack of malaria (16). In this patient, a highlevel of parasitemia contrasted with the absence of any severeclinical symptoms related to sequestration and peripheral bloodwas shown to contain not only the ring forms but all differentstages of parasitic development. These peculiar characteristicsof the infection suggest that, as in splenectomized monkeys,parasite sequestration was reduced in this splenectomized pa-tient.

Knobs have been shown to mediate the binding of tropho-zoite/ and schizont-infected erythrocytes to the vascular en-dothelium (3, 4). Although infected erythrocytes from S- an-imals fail to bind in vitro and to be sequestered in vivo, both

washing, melanoma cells covered with infected erythrocytes were in-cubated for 1 hr with a 1:20 dilution of serum 31/S+ or of normal serum.51Cr in the culture supernatant was then measured and related to 50Crbound to the melanoma cells before incubation with serum.

Proc. Natl. Acad. Sci. USA 80 (1983)

60

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 3,

202

0

Proc. Natl. Acad. Sci. USA 80 (1983) 5079

infected erythrocytes from S' and S monkeys have knobs thatare indistinguishable in morphology and numbers. This indi-cates that an ultrastructural description is not sufficient to de-fine a functional binding site. P. falciparum in S monkeys couldbe comparable with other species of malaria such as Plasmo-dium malariae, which produces knobs but is not sequestered(17).The study of in vitro binding with enzyme-treated eryth-

rocytes provides some information on the nature of the surfacemolecules involved in sequestration. These molecules are as-sociated at least in part with proteins of the infected erythro-cyte membrane since binding is inhibited by treatment of in-fected blood with proteases. Proteases are only effective onerythrocytes containing parasites that have reached a certaindegree of maturation. If protease treatment is applied earlier-i.e., at the ring stage, it has no effect on the results of the MCBAcarried out once the parasites have reached a sufficient degreeof maturation. This indicates either that the protease-sensitivemolecules are of parasitic origin and are not yet exposed on theerythrocyte membrane at the ring stage or that, if a host com-ponent is involved, it becomes sensitive to protease treatmentonly after being first modified or exposed by parasite devel-opment. Increased binding of neuraminidase-treated eryth-rocytes infected with parasites from S' animals to melanomacells could be related to a decrease in negative cell surface chargeor to the exposure of a new carbohydrate residue involved inbinding. The latter hypothesis assumes the presence of a lectin-like molecule on the surface of melanoma cells; such lectinshave been described on various types of mammalian cells (18,19). The increased reactivity of P. falciparum-infected squirrelmonkey erythrocytes with certain lectins has previously beendemonstrated (20), but the relationship between binding prop-erties to lectins and to melanoma cells has not yet been found.We have shown that antibody had an effect on the cytoad-

herence of infected erythrocytes in vitro and in vivo. The in-hibition of in vitro binding of infected erythrocytes by immuneserum indicates that antibody can interact with the structuresof the erythrocyte that mediate binding. This interaction couldbe direct, through a mechanism of competitive inhibition, orindirect, through an antibody-mediated surface or cytoskeletalalteration of the membrane environment that inactivated thebinding site. In a study performed in parallel, Hommel et aL(11) have shown by immunofluorescence that the surface of in-fected erythrocytes from S' animals reacted only with serum31/S' and that the surface of infected erythrocytes from S5animals reacted only with serum 30/S-. The low inhibitory ef-fect on binding of serum 30/S-, when compared with serum31/S+, suggests a close linkage between the binding propertiesof the infected erythrocyte surface and the surface antigens de-tected on the infected erythrocyte by immunofluorescence.

In vivo, the sharp increase of circulating trophozoites/schiz-onts that occurs in an intact monkey within minutes after pas-sive transfer of immune serum clearly indicates that the re-versal of binding can occur in vivo as well as in vitro. Thesecondary decrease in parasitemia that is observed 24 hr latersupports the view that an interaction of antibody with seques-tration might represent the initial step of a chain of events thatwill eventually destroy the parasite. In the light of this inhib-itory effect of immune serum on binding, it is of interest to notethat the prevalence of cerebral malaria has been shown to be

higher in children and in nonimmune individuals newly ex-posed to malaria (21). In some context, the "crisis forms" (i.e.,deteriorating trophozoites and schizonts) that appear in thebloodstream of infected hosts at the onset of an effective im-mune response, just before the immune system finally resolvesthe infection (22, 23), might represent parasites whose bindingto endothelial cells has been inhibited by circulating antibodiesand that have subsequently been killed as a consequence of theirpresence in the peripheral bloodstream. Such observations couldbe related to the existence of antibody in the serum of immuneindividuals that would inhibit parasite sequestration, thus al-leviating the clinical severity of the disease. The biochemicalnature of the molecules present on the surface of the infectederythrocyte that mediate sequestration is not yet known, but itis conceivable that such molecules might constitute a potentialimmunogen.

We wish to thank Professor J. R. David for continuous encourage-ment and helpful discussions at every stage of this study and Dr. I. Greenfor important suggestions on the manuscript. This work was supportedby the United Nations Development Program/World Bank/WorldHealth Organization Special Programme for Research and Training inTropical Diseases and by the Rockefeller Foundation.

1. Garnham, P. C. C. (1966) in Malaria Parasites and Other Hae-mosporidia (Blackwell, Oxford), pp. 357-430.

2. Miller, L. H. (1969) Am. J. Trop. Med. Hyg. 18, 860-865.3. Trager, W., Rudzinska, M. A. & Bradbury, P. C. (1966) Bull.

W. H. 0. 35, 883-885.4. Luse, S. A. & Miller, L. H. (1971) Am. J. Trop. Med. Hyg. 20, 655-

660.5. Kilejian, A., Abati, A. & Trager, W (1977) Exp. Parasitol 42, 157-

164.6. Langreth, S. G. & Reese, R. T. (1979)J. Exp. Med. 150, 1241-1254.7. Udeinya, I. J., Schmidt, J. A., Aikawa, M., Miller, L. H. & Green,

I. (1981) Science 213, 555-557.8. Schmidt, J. A., Udeinya, I. J., Leech, J. H., Hay, R. J., Aikawa,

M., Barnwell, J., Green, I. & Miller, L. H. (1982)J. Clin. Invest.70, 379-386.

9. Kreier, J. P. & Green, T. J. (1980) in Malaria, ed. Kreier, J. P.(Academic, New York), Vol. 3, pp. 111-162.

10. Gysin, J., Hommel, M. & Pereira Da Silva, L. (1980)J. Parasitol.66, 1003-1013.

11. Hommel, M., David, P. H. & Oligino, L. D. (1983)J. Exp. Med.,157, 1137-1148.

12. Diggs, C. L., Joseph, K., Flemmins, B., Snodgrass, R. & Hines,F. (1975) Am. J. Trop. Med. Hyg. 24, 760-763.

13. Trager, W. & Jensen, J. (1976) Science 193, 673-675.14. Hommel, M., David, P. H., Oligino, L. D. & David, J. R. (1982)

Parasite Immunol. 4, 409-419.15. Barnwell, J. W., Howard, R. J., Coon, H. G. & Miller, L. H. (1983)

Infect. Immun., in press.16. Garnham, P. C. C. (1970) Acta Trop. 27, 1-4.17. Smith, D. H. & Theakston, R. D. G. (1970) Ann. Trop. Med. Par-

asitol 64, 329.18. Ashwell, G. & Morell, A. G. (1977) Trends Biochem. Sci. 2, 76-

78.19. Kieda, C. M. T., Bowles, D. J., Ravid, A. & Sharon, N. (1978)

FEBS Lett. 94, 391-396.20. David, P. H., Hommel, M. & Oligino, L. D. (1981) Mol Biochem.

Parasitol. 4, 195-204.21. Gelfand, M. (1973) in Tropical Neurology, ed. Spillane, J. D.

(Oxford Univ. Press, London), pp. 247-258.22. Taliaferro, W. H. & Taliaferro, L. G. (1944)J. Infect. Dis. 80, 78-

104.23. Barnwell, J. W. & Desowitz, R. S. (1977) Ann. Trop. Med. Par-

asitol. 71, 429-433.

Medical Sciences: David et al.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 3,

202

0